test_async_llm.py 8.73 KB
Newer Older
1
2
# SPDX-License-Identifier: Apache-2.0

3
import asyncio
4
from contextlib import ExitStack
5
from typing import Optional
6
7
8
9

import pytest

from vllm import SamplingParams
10
from vllm.assets.image import ImageAsset
11
from vllm.engine.arg_utils import AsyncEngineArgs
12
from vllm.inputs import PromptType
13
from vllm.platforms import current_platform
14
from vllm.sampling_params import RequestOutputKind
15
16
17
18
19
20
from vllm.v1.engine.async_llm import AsyncLLM

if not current_platform.is_cuda():
    pytest.skip(reason="V1 currently only supported on CUDA.",
                allow_module_level=True)

21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
TEXT_ENGINE_ARGS = AsyncEngineArgs(model="meta-llama/Llama-3.2-1B-Instruct",
                                   enforce_eager=True,
                                   disable_log_requests=True)

VISION_ENGINE_ARGS = AsyncEngineArgs(model="Qwen/Qwen2-VL-2B-Instruct",
                                     enforce_eager=True,
                                     disable_log_requests=True)

TEXT_PROMPT = "Hello my name is Robert and"

VISION_PROMPT_TEMPLATE = (
    "<|im_start|>system\nYou are a helpful assistant.<|im_end|>"
    "\n<|im_start|>user\n<|vision_start|><|image_pad|><|vision_end|>"
    "What is in the image?<|im_end|>\n"
    "<|im_start|>assistant\n")
VISION_PROMPT = {
    "prompt": VISION_PROMPT_TEMPLATE,
    "multi_modal_data": {
        "image": ImageAsset("stop_sign").pil_image
    }
}
42
43


44
45
async def generate(engine: AsyncLLM,
                   request_id: str,
46
                   prompt: PromptType,
47
                   output_kind: RequestOutputKind,
48
                   max_tokens: int,
49
                   n: int = 1,
50
                   prompt_logprobs: Optional[int] = None) -> tuple[int, str]:
51
52
53
    # Ensure generate doesn't complete too fast for cancellation test.
    await asyncio.sleep(0.2)

54
    count = 0
55
    sampling_params = SamplingParams(max_tokens=max_tokens,
56
                                     ignore_eos=True,
57
                                     output_kind=output_kind,
58
59
60
                                     temperature=0.5,
                                     seed=33,
                                     n=n,
61
                                     prompt_logprobs=prompt_logprobs)
62
    async for out in engine.generate(request_id=request_id,
63
                                     prompt=prompt,
64
65
                                     sampling_params=sampling_params):

66
        num_tokens = sum(len(output.token_ids) for output in out.outputs)
67
68
69
70
        if output_kind == RequestOutputKind.DELTA:
            count += num_tokens
        else:
            count = num_tokens
71
72
73
74
75
76

        await asyncio.sleep(0.)

    return count, request_id


77
78
@pytest.mark.parametrize(
    "output_kind", [RequestOutputKind.DELTA, RequestOutputKind.FINAL_ONLY])
79
80
81
@pytest.mark.parametrize("engine_args_and_prompt",
                         [(TEXT_ENGINE_ARGS, TEXT_PROMPT),
                          (VISION_ENGINE_ARGS, VISION_PROMPT)])
82
@pytest.mark.asyncio
83
84
85
86
87
async def test_load(
    monkeypatch: pytest.MonkeyPatch,
    output_kind: RequestOutputKind,
    engine_args_and_prompt: tuple[AsyncEngineArgs, PromptType],
):
88
89
90
    # TODO(rickyx): Remove monkeypatch once we have a better way to test V1
    # so that in the future when we switch, we don't have to change all the
    # tests.
91
    with monkeypatch.context() as m, ExitStack() as after:
92
        m.setenv("VLLM_USE_V1", "1")
93
        engine_args, prompt = engine_args_and_prompt
94

95
        engine = AsyncLLM.from_engine_args(engine_args)
96
        after.callback(engine.shutdown)
97

98
        NUM_REQUESTS = 100
99
100
101
102
103
104
105
106
107
        NUM_EXPECTED_TOKENS = 10

        request_ids = [f"request-{i}" for i in range(NUM_REQUESTS)]

        # Create concurrent requests.
        tasks = []
        for request_id in request_ids:
            tasks.append(
                asyncio.create_task(
108
                    generate(engine, request_id, prompt, output_kind,
109
                             NUM_EXPECTED_TOKENS)))
110
111

        # Confirm that we got all the EXPECTED tokens from the requests.
112
113
114
115
116
        done, pending = await asyncio.wait(tasks,
                                           return_when=asyncio.FIRST_EXCEPTION)
        for task in pending:
            task.cancel()
        for task in done:
117
            num_generated_tokens, request_id = await task
118
119
120
121
122
123
124
            assert num_generated_tokens == NUM_EXPECTED_TOKENS, (
                f"{request_id} generated {num_generated_tokens} but "
                f"expected {NUM_EXPECTED_TOKENS}")

        assert not engine.output_processor.has_unfinished_requests()


125
126
@pytest.mark.parametrize(
    "output_kind", [RequestOutputKind.DELTA, RequestOutputKind.FINAL_ONLY])
127
128
129
@pytest.mark.parametrize("engine_args_and_prompt",
                         [(TEXT_ENGINE_ARGS, TEXT_PROMPT),
                          (VISION_ENGINE_ARGS, VISION_PROMPT)])
130
@pytest.mark.asyncio
131
132
async def test_abort(monkeypatch: pytest.MonkeyPatch,
                     output_kind: RequestOutputKind,
133
                     engine_args_and_prompt: tuple[AsyncEngineArgs,
134
                                                   PromptType]):
135

136
    with monkeypatch.context() as m, ExitStack() as after:
137
        m.setenv("VLLM_USE_V1", "1")
138
        engine_args, prompt = engine_args_and_prompt
139

140
        engine = AsyncLLM.from_engine_args(engine_args)
141
        after.callback(engine.shutdown)
142
143
144

        NUM_REQUESTS = 100
        NUM_EXPECTED_TOKENS = 100
145
        NUM_EXPECTED_TOKENS_LONG = 50000
146
        REQUEST_IDS_TO_ABORT = range(1, 100, 10)
147
        PARALLEL_SAMPLE_REQ_IDS = range(1, 100, 15)
148
149
150
151

        request_ids = [f"request-{i}" for i in range(NUM_REQUESTS)]

        # Create concurrent requests.
152
        tasks: list[asyncio.Task] = []
153
154
155
156
        for idx, request_id in enumerate(request_ids):
            max_tokens = NUM_EXPECTED_TOKENS_LONG if (
                idx in REQUEST_IDS_TO_ABORT) else NUM_EXPECTED_TOKENS
            n = 3 if idx in PARALLEL_SAMPLE_REQ_IDS else 1
157
158
            tasks.append(
                asyncio.create_task(
159
                    generate(engine, request_id, prompt, output_kind,
160
                             max_tokens, n)))
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175

        # API server cancels requests when they disconnect.
        for idx in REQUEST_IDS_TO_ABORT:
            tasks[idx].cancel()
            await asyncio.sleep(0.1)

        # Confirm the other requests are okay.
        for idx, task in enumerate(tasks):
            # Confirm that it was actually canceled.
            if idx in REQUEST_IDS_TO_ABORT:
                with pytest.raises(asyncio.CancelledError):
                    await task
            else:
                # Otherwise, make sure the request was not impacted.
                num_generated_tokens, request_id = await task
176
177
178
                n = 3 if idx in PARALLEL_SAMPLE_REQ_IDS else 1
                expected_tokens = NUM_EXPECTED_TOKENS * n
                assert num_generated_tokens == expected_tokens, (
179
                    f"{request_id} generated {num_generated_tokens} but "
180
                    f"expected {expected_tokens}")
181

182
        # Make sure all aborted requests were really aborted.
183
184
185
186
187
        assert not engine.output_processor.has_unfinished_requests()

        # Confirm we can do another generation.
        request_id = f"request-{REQUEST_IDS_TO_ABORT[0]}"
        task = asyncio.create_task(
188
189
            generate(engine, request_id, prompt, output_kind,
                     NUM_EXPECTED_TOKENS))
190
191
192
        num_generated_tokens, request_id = await task
        assert num_generated_tokens == NUM_EXPECTED_TOKENS
        assert not engine.output_processor.has_unfinished_requests()
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225


@pytest.mark.parametrize("n", [1, 3])
@pytest.mark.parametrize("engine_args_and_prompt",
                         [(TEXT_ENGINE_ARGS, TEXT_PROMPT),
                          (VISION_ENGINE_ARGS, VISION_PROMPT)])
@pytest.mark.asyncio
async def test_finished_flag(monkeypatch, n: int,
                             engine_args_and_prompt: tuple[AsyncEngineArgs,
                                                           PromptType]):

    with monkeypatch.context() as m, ExitStack() as after:
        m.setenv("VLLM_USE_V1", "1")
        engine_args, prompt = engine_args_and_prompt

        engine = AsyncLLM.from_engine_args(engine_args)
        after.callback(engine.shutdown)

        sampling_params = SamplingParams(max_tokens=100,
                                         output_kind=RequestOutputKind.DELTA,
                                         temperature=1.0,
                                         seed=33,
                                         n=n)
        outputs = [
            out
            async for out in engine.generate(request_id="request-33",
                                             prompt=prompt,
                                             sampling_params=sampling_params)
        ]

        # Assert only the last output has the finished flag set
        assert all(not out.finished for out in outputs[:-1])
        assert outputs[-1].finished